Fundamental Processes in the Photodegradation of Polymers

19 Fundamental Processes in the Photodegradation of Polymers J. E. GUILLET, J. DHANRAJ, F. J. GOLEMBA, and G. H. HARTLEY

Downloaded by YORK UNIV on October 19, 2014 | http://pubs.acs.org Publication Date: June 1, 1968 | doi: 10.1021/ba-1968-0085.ch019

University of Toronto, Toronto 5, Canada A study has been made of the mechanism of some of the more fundamental photochemical processes which lead ultimately to polymer degradation. By locating ketone chromophore groups in different positions within the polymeric molecule, the relationship between molecular structure and the quantum efficiency for various modes of energy dissipation have been established. This study, along with studies on the effects of molecular weight, temperature, and internal viscosity, indicates that the major cause of degradation in polymers containing ketone groups is the Norrish Type II photoelimination. The quantum efficiency of the reaction appears to be nearly independent of temperature and of the physical state of the polymer. At normal temperatures the quantum efficiency for radical processes is quite low. The mode of action of two types of stabilizers is briefly described.

*"|^he deterioration of polymers in outdoor applications is usually caused by a complex series of reactions, initiated by the absorption of ultraviolet light. This paper discusses the nature of certain primary photochemical reactions which lead ultimately to degradation and establishes how they are affected by experimental conditions. Since photochemical reactions of relatively small organic molecules have been studied extensivly, it would also be of interest to be able to predict the photochemistry of polymer molecules from the known chemistry of simpler compounds. The effect of various types of radiation on polymers has been studied extensively; however, if we are interested mainly in effects which occur in a terrestrial atmosphere (i.e., conventional "wea&ering") we must restrict our attention to a rather narrow band of electromagnetic spectrum. In fact, we must concern ourselves primarily with the energy emitted by the sun and transmitted by the earth's atmosphere. 272

In Stabilization of Polymers and Stabilizer Processes; Platzer, N.; Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

19.

GuiLLET E T A L .

273

Photodegradation of Polymers


Photochemical Considerations The energy distribution of solar radiation as a function of wavelength ( λ ) is shown in Figure 1, Curve A (11). This is the energy which would reach an object on the surface of the earth if there were no atmosphere to absorb part of the radiation. Curves Β and C represent the energy distribution reaching the earth after passing through the atmosphere at high noon and when the angle of the sun is 30° to the horizontal, respec tively. It is readily apparent that no radiation with a wavelength shorter than 3000 A . is transmitted by the atmosphere. This fact has important practical implications i n that the number and type of photochemical reactions which may occur are thereby severely restricted. Since the energy of light is inversely proportional to the wavelength (i.e., E\ = he/λ), the removal of the short wavelength radiation means that only the relatively low energy light reaches the earths surface. If, for example,

0

5

10

15

20

WAVELENGTH ( A x 10" ) 3

Figure 1.

Energy distribution of sohr radiation

we are concerned with processes involving the breaking of chemical bonds, radiation w i l l be available to break only the weakest of such bonds. This is shown in Figure 2 where the proportion of the sun's total energy greater than hc/k is plotted as a function of Ex. Values of the strengths of several chemical bonds are included for reference. Although over half of the sun's radiation has sufficient energy to break weak bonds such as Ο—Ο or Ν—Ν, only about 5 % of the total is sufficiently energetic to break a carbon-carbon bond, and none is expected to break strong bonds such as C—Η, Ο—H, C = C , and C = 0 .


274

STABILIZATION OF POLYMERS A N D STABILIZER PROCESSES


A further restriction on the possible reactions is imposed b y the elementary photochemical principle that light must be absorbed by the polymer if a reaction is to occur. However, most pure synthetic polymers do not absorb at wavelengths longer than 3000 A . and hence should not be affected by terrestrial ultraviolet light. This is demonstrated in Figure 3 which shows the absorption spectra of two typical vinyl polymers.

Figure 2.

Distribution of sohr energy with bond strengths of simple chemical bonds

However, these polymers do degrade when subjected to terrestrial ultra violet radiation, and this has been attributed to the presence of small amounts of impurities which absorb light and initiate oxidative chain reactions within the polymer. In the oxidation process, compounds con taining peroxy and keto groups are formed, and these absorb in the long wavelength region and accelerate the breakdown of the polymer chains. Obviously a detailed study of such reactions i n polymers w i l l be difficult because of the frequently unknown nature of the trace impurities which initiate the reactions and of the multiplicity of products formed i n the photooxidation. In the present studies this difficulty has been eliminated by synthe sizing polymers with well-defined chromophoric groups situated at known locations with respect to the polymer chains. B y studying the photoreactions of these polymers in comparison with those of suitable model compounds, it was possible to infer much about the photochemistry of macromolecules.


19.

GuiLLET E T A L .

275


Experimental Two light sources were used; one was a 1-kw. high pressure (110 atm.) water-cooled mercury arc, type A H - 6 (General Electric); the other was a 250-watt, medium pressure (30 atm.) compact source mercury arc, type M E / D (Associated Electrical Industries L t d . ) . The 3130-A. line was isolated by a sheet of borosilicate glass plate and 2 cm. of filter solution containing nickel and cobalt sulfate. Analysis of the light from the M E / D lamp filtered by this system showed a main peak at 3130 Α., with smaller peaks at 2900 and 3360 A . The light intensity from the lamp was monitored by a 1P28 photomultiplier tube, powered by a Lambda


POLY (METHYL METHACRYLATE )

10 POLY(STYRENE)

0-5

LU

ο

ο

-0-5

-10

320

300

280

260

240

220

WAVELENGTH (πημ) Figure 3.

Absorption spectra of polystyrene and poly(methyl methacrylate)


276


model 24 regulated power supply; the signal from the phototube was measured on a L & Ν model G recorder. The intensity of the light was measured, and the photomultiplier tube was calibrated using uranyl oxalate actinometry. The response of the photomultiplier tube was linear over the range of intensities measured (0-8.5 Χ 10" Einst./hr.). The light from the lamp was focused by a quartz lens into a parallel beam, which passed successively through a filter cell, a bath containing the reaction cell, and into the photomultiplier tube. The filter cell, the reaction cell, and the windows of the bath were all of fused quartz. The lens, bath, and photomultiplier tube housing were rigidly mounted on an optical bench. Temperature control inside the bath was achieved by using either a circulating water pump ( ±:0.25°C.) or a hot air blower ( ± r l ° C . ) . For photolysis of the ethylene-carbon monoxide copolymer i n solu tion, the A H - 6 lamp and a 20-mm. path-length quartz cell were used. The cell was filled with the solvent, pure η-heptane, and the intensity of the lamp was measured at the experimental temperature. Freeze-dried polymer was then added to make a 2 % solution, which absorbed about 25% of the light. The polymer was dissolved, and the solution was mixed by a dry nitrogen stream, which also flushed out any air dissolved in the solvent. The light beam was then allowed to enter the cell, and the pho tolysis commenced; the intensity of the emergent beam was monitored by the photomultiplier tube and the recorder. A t the end of the photolysis the cell was filled with pure solvent, and the intensity of the lamp was measured again. The polymer was recovered from solution by evaporat ing the heptane; it was then dissolved in benzene and freeze-dried. The number average molecular weight of the degraded polymers was measured on an ebulliometer, using toluene as a solvent and a tristearin standard. Intrinsic viscosities of the polymers were determined using a Desreux dilution viscometer. Infrared spectra of the polymer, before and after photolysis, were run on a Perkin Elmer 521 double-beam grating spectrophotometer, using compression-molded films. F i l m thicknesses were determined by a micrometer. The solvents used in the photolyses were purified by distillation and filtered through silica gel columns until they were completely transparent over the range of wavelengths passed by the filter system. A Perkin Elmer model 800 gas chromatograph with flame ionization detectors and 1/8-inch χ 5-foot columns, 5 % w / w silicone SE30 on chromosorb G , was used to analyze the photolysis products. Ultraviolet spectra were run on a Bausch and L o m b Spectronic 505 double-beam spectrophotometer. Carbon monoxide, a product of the polymer and ketone photolyses, was collected and measured on a gas chromatograph, in a separate series of experiments. The polymer solution was flushed with hydrogen, before, during, and after photolysis, sweeping any evolved gases through acti vated charcoal cooled in liquid nitrogen. The tube containing the char coal was then closed, most of the hydrogen was pumped off at liquid nitrogen temperature, and the tube was connected to the gas chromato graph. After the tube warmed to room temperature, any desorbed gases were swept into the gas chromatograph by a hydrogen stream (carrier gas), where they were separated by a 5-A. molecular sieve column using


4



19.

GuiLLET E T A L .

277


a thermal conductivity detector. The peak areas were compared with the peak areas of known volumes injected before and after the collected sample. The 1 % ethylene-carbon monoxide copolymer was also irradiated in the solid phase (thin film). Compression-molded films were fixed on plates which fitted into the Perkin Elmer 521 infrared spectrophotometer. A n infrared spectrum of the polymer could thus be obtained after each period of photolysis without disturbing the film. For photolyses at room temperature and above the plates were mounted in a solid brass cell through which a stream of inert gas could be passed while the cell was being heated. The ketones used were i n most cases obtained from commercial sources and purified by distillation. Analysis by gas chromatography showed a purity greater than 9 8 % i n all cases. D i a l k y l ketones not avail able commercially were synthesized from the appropriate acids using an iron catalyst and the method of Davis and Schultz ( 3 ) . Results and Discussion Photochemistry of Ketones. F r o m the point of view of weathering damage the most important of the chromophores absorbing i n the near ultraviolet w i l l be those containing oxygen since they are most likely to be formed i n polymers degraded i n an oxygen atmosphere. The group which has been studied most extensively is the ketone group, C = 0 . This shows a relatively weak absorption maximum i n the range of 270-290 m/x. This is caused by the transition when an electron is raised from a nonbonding η orbital localized on the oxygen atom into a delocalized antibonding π * orbital which extends over the entire carbonyl group. Using Kasha's notation this is an η-ττ* transition, and is always the electronic transition of lowest energy i n carbonyl and nitroso compounds. It is characteristic of groups containing "lone pair" electrons such as C = 0 , C = S , — N = 0 , — N 0 , and — Ο — N = 0 . Some vibrational structure can be seen i n the vapor state, but i n solution the absorption is essentially continuous. 2

One of the simplest polymers containing the ketone group can be synthesized by the high pressure copolymerization of ethylene and carbon monoxide: Ο

CHo=CH + CO —> CHoCHo L(CH CH ) C— (CH CH ) 2

2

2

n

2

2

n

,CH=CH

9

where η can be any integer, and χ is a large number. The structure approximates that of a linear hydrocarbon chain containing ketone groups at random intervals, determined by the mole ratio of carbon monoxide used i n the polymerization. If the polymer is formed at high tempera tures, it w i l l also contain short branches (ca. four carbon atoms), and if


278

STABILIZATION O F POLYMERS A N D STABILIZER PROCESSES

the concentration of C O is small, the physical properties of the polymer are almost indistinguishable from those of "high pressure" polyethylene. The ultraviolet absorption spectrum of such a polymer (containing 1 mole % C O ) is compared i n Figure 4 with that of a typical model ketone ( 12-tricosanone ). 1

1

1

1

1

1

1

20

N

ω

10

—

^ s E t - C o Copolymer

/ /

Log


1-5

/ /

0-5

/

/

\

/

\ \

/

\

/

\

12-Tricosanone

Λ U

0 CHj(CH )| C(CH ) CH 2

200

1 220

1 240

1 260

1 280

1 300

0

1 '320

2

l0

s

1 340

WAVELENGTH (ιτιμ) Figure 4. Absorption spectrum of ethylene-carbon monoxide (1%) co polymer in heptane at 80°C. The molar extinction coefficient (c) and A x for the polymer and the model compound are almost identical, and the spectra are sufficiently similar to suggest that the nature of the absorption process is the same in both cases. When a molecule absorbs a quantum of radiation, it is raised to an excited state. A tfirstthe energy w i l l be localized, usually in the electronic energy of a particular electron associated with a chromophoric group. The quantum of energy so absorbed may be dissipated by either photophysical or photochemical processes such as those listed below: ma

Photophysical Fluorescence Phosphorescence Thermal Energy Transfer

Photochemical Photoionization Free Radical Formation Cyclization Intramolecular Rearrangement



19.


GUILLET ET A L .

279

Since most chemical reactions if carried to completion w i l l result in deterioration of polymer properties, it is desirable to be able to ensure that a l l of the energy is dissipated in photophysical processes and to eliminate the photochemical reactions. However, i n some cases, such as in the manufacture of photographic resists, it is desirable to maximize the photochemical effects. The polymer chemist is particularly concerned with the problem of how the relative efficiency of these various processes may be affected by the polymeric nature of the molecules he uses. It is convenient for photochemical studies i n the near ultraviolet to use the mercury resonance line at 3130 A . from a medium pressure mer cury arc as a light source. This particular wavelength can be separated in high intensity from the remainder of the spectrum b y suitable filters. Light quanta of this wavelength have an energy equivalent to 91.4 kcal./mole. This is less than the strength of the carbon-oxygen bond i n the carbonyl group but more than enough to break a carbon-carbon single bond. F o r example, for acetone the following reaction: Ο

Ο

C H j—d)—CHo

*CH

3

•+ -CH

(1)

3

requires about 71 kcal./mole. If this is the reaction path, the two frag ments would retain an additional 20.4 kcal. as thermal energy. Norrish and co-workers i n a series of early papers (2) showed that the chemical products of decomposition of aliphatic ketones could be explained on the basis of two primary reactions, the first designated Type I (Reaction 1)—a homolytic split of the molecule to give two free radicals— and Type II, which gives an olefin and a methyl ketone

(Reaction^)

Ο

Ο h* Il C H C H C H — C — C H -> C H = C H + C H — C — C H (2) This step does not appear to proceed b y a free radical mechanism and often has about the same quantum efficiency at all temperatures. In addition to these, a third reaction has been identified recently (15) which results in the formation of a cyclobutanol derivative: Il

3

2

2

3

2

2

Ο κ

II

CH CH2CH —C—CH 3

2

3

3

OH ι * CH —C—CH 2

3

3

(3)

C H 2—^H

2

This process usually has lower quantum efficiency than the Type II reaction and is less likely to occur i n "the higher ketones.


280


Polymer Photochemistry. The occurrence of these reactions in poly meric ketones was first demonstrated by Guillet and Norrish (6, 7), who studied poly (methyl vinyl ketone) in solution and showed that the main features of the photodegradation could be accounted for quantitatively on the basis of Type I and Type II reactions. The conclusion was later confirmed by Wissbrun (13). Recent studies of the ethylene-carbon monoxide polymer (9) confirm that both Type I and Type II reactions occur. The Type I reaction results in the formation of two polymer radi cals, one of which is an acyl radical which may subsequently decarbonylate (Reaction 4).


Ο ' CH2CH2CH2CH2CH2CH2—C—CH2CH2CH2CH2CH2CHJ2

ih

O

v

(4)

—' CH2CH2CH2CH2CH0CH2" -t~ 'CCr^Cf^CrLjCH^CI^CrLj CO +

CHaCHaCHaCHaCHaCHa —

The Type II reaction also leads to a break i n the "backbone" of the polymer chain with the formation of a methyl ketone and a terminal vinyl group (Reaction 5). Ο —CHaCrlaCrlaCrLjCI^CI^—C—CH2CrIaCrl2CHaCrIaCH2— i

k

v

M

(5)

— C r L j C r ^ C r L j C r l = CHa ~\~ C H 3 — C — C H a C H a C H o C r l a C H a C r l a —

A t 25 °C. the predominant reaction is Type II, and it is the primary cause of the decrease in molecular weight which occurs on irradiation. A t higher temperatures Type I becomes more favorable, and the decrease in molecular weight is slowed down by a "repolymerization" caused by addition of the polymer radicals to the double bonds formed by the Type II reaction (7). The Type II reaction has been shown to cause polymer degradation not only in poly (methyl vinyl ketone) and poly (ethylene— C O ) but in polyesters such as polyethylene terephthalate (12). It has been suggested that i t may be the major cause of degradation in a wide variety of synthetic polymers containing ester or keto carbonyl groups (12). Effect of Molecular Weight. Studies of the variation in the quantum efficiencies of various primary photochemical processes i n polymers showed very little dependence on molecular weight. Accordingly, model compounds of the general structure

Ο

|| were studied (9). R—C—R


The

19.

GuiLLET E T A L .

281


results are shown i n Figure 5 i n which the quantum efficiency for the Type I reaction (estimated from the yield of C O ) at 120°C. ( φ ι ) is plotted as a function of total chain length. There is an abrupt decrease in φ ι for ketones with chain length greater than seven (i.e., R = C ) , after which φ ι decreases gradually with chain length to a limiting value of .012. The value for poly (ethylene—CO) is about twice this limiting value, which suggests strongly that some of the C O groups are i n side branches rather than i n the main chains. Approximately the same value of φ ι ( 0.018 ) was obtained when the polymer was irradiated i n film form at 120 °C. i n a hydrogen atmosphere. 3


0-40

Ί

1

1

Γ

1

0-30, 003h

CO 002|-

O'Olh

0

10

J

20

ι

L

30

40

CHAIN LENGTH Figure 5.

Quantum yields for carbon monoxide formation in ketones of type Ο R—C—R,

determined in hydrocarbon solvent at 120°C.

Similar data on methyl ketones ( I I )

Γ

° ||

Ί

at 25°C, are

LR_C—CH J 3

shown i n Table I. Again there is a sharp decrease i n φι after R = C . However, the value of φι for poly (methyl vinyl ketone) is considerably higher than that of any of the model ketones probably because of the lower energy required to form a secondary rather than a primary radical. The value of φ varies from 0.11 to 0.22 i n both types of model ketones and is almost independent of chain length over the range studied. O n 3

π


282


the other hand, both poly (methyl vinyl ketone) and poly (ethylene—CO) have values of φ = 0.025, and this value appears to be independent of temperature, and more surprising, it is the same whether the polymer is in solution or i n film form. If one accepts the cyclo-elimination mecha nism (Figure 6) for the Type II reaction, the excited state involved must have a sufficiently long lifetime so that large changes i n the mobility of the polymer chain do not affect the probability of reaction. η

Table I.

Quantum Efficiencies at 25°C. for Methyl Ketones Ο [ R — C — C H ] in Hydrocarbon Solution


3

α 5 5 5 7

0.21 0.21 0.21 0.22 0.025

0.023 0.003 0.0022 0.0003 0.04

Poly ( methyl vinyl ketone )

Réf.

φπ

Φι

"Estimated from data of Ausloos and Rebbert (1).

ïï '

X

CH

N

2

CH

\CH

2

CH

2

\CH

2

2

htf

CH, '

CH CH

N

2

2 N

/ CH -CH 2

N

H

" " ° ^ / ^C—CH

C

H

z

2

CH -CH 2

V XH

/

C

H

2 \ N

2

2

I '

X

CHi

JZ — C H

\CH -CH 2

^CH

2

CH

X

2

CH

\

2

2

I

Fundamental Processes in the Photodegradation of Polymers

Recommend Documents